Many attempts have been made to associate dental caries in man with antibodies in serum or saliva. The results from studies of adult saliva are unclear but salivary IgA antibodies to S. mutans have been demonstrated to experimental caries in the rodent model. Recently however it was reported that children with high levels of antibody to S. mutans in their saliva had no detectable caries. Studies of serum antibodies in man have shown that natural serum antibodies to S. mutans may be protective, and that subjects who are able to maintain these antibody levels for many years are naturally protected. These studies have been valuable contributions to our understanding coronal caries and could prove equally valuable in the investigation of the etiology and prevention of root caries. The present study should indicate not only the presence of protective antibody in serum or saliva, but specificity to an organism which may influence future microbiological investigations. The patients for this study will be 65 and over and drawn from a large population already identified and available to the investigators. Fifty patients in each of 5 groups will be studied and will include those with and without root caries, with and without exposed root surfaces and subjects with no standing teeth. Serum and saliva samples will be collected at 6 month intervals, with a group of 10 patients undergoing a more intensive six month investigation to assess stability of local levels of specific antibodies and immunoglobulins. Assays for antibodies and their specific activity will be carried out using a new version of ELISA assays. Antibodies against whole cells of bacteria normally found in root surface plaque will be assayed. These bacteria include, S. mutans, S. milleri, A. viscosus, and A. naeslundii, S. sanguis, Lactobaccillus casae, and E. coli (control). Plaque samples will be ultrasonicated, smeared, air dried and identified using rhodominated or fluorescinated antisera to individual bacteria. Should some meaningful immunological correlate be identified, isolation of antigens specific for the bacteria of interest is a natural second step.